Per TIP-92, immediate withdrawal applies only to non-beta stakers. The stake held by beta stakers will not be withdrawable until the beta nodes transition to DAO-provided delegations or a DAO-controlled whitelist. There are 6 beta staker nodes operated by Boar. Those nodes were being compensated entirely via rewards and operationally paid for by other entities out of pocket. All those stakers are early supporters of the network participating in the very first signing groups.
Out of all currently operating tBTC wallets, up to 21% of group key shares are held on those nodes. Although with a 51-of-100 threshold disabling those nodes should not kill the liveness of wallets, in practice, the outcome will depend on the rest of the peers and the 21% liveness reduction is quite significant. Disabling those nodes today significantly increases the risk to the network with up to 21% liveness reduction of signing groups.
With TIP-92 changes, the company is put in an extremely difficult position. The stakes of customers will remain locked but they will not receive 15% rewards for their nodes as initially agreed.
Also, the fact beta stakers remain locked is something new, introduced in TIP-92. Before TIP-92, the beta stakers were required to cooperate with the DAO on unstaking and transitioning but their stake was not locked.
To address this issue we propose to allow unlocking the stake of those beta stakers immediately, just like the rest of the stakers. In addition to that, we commit to keep running those nodes for $2000/mo per node, paid in T starting Feb 14th, with an option to switch to a single node in the future with the DAO delegation. Although nodes with no stake will not participate in key generation rituals, they will continue to respond to signing requests. The proposed solution fixes TIP-92 rugging a selected group of stakers and prevents degradation of the network.
Note that these were the first beta staker nodes to take on all the risk and effort at the launch of the tBTC v2 protocol, and now they are being left rugged. This proposal addresses the unacceptable clutter and confusion caused by the recent other TIPs. The problem has to be solved ASAP as it’s crucial for network stability.
Boar was one of the three initial staking providers that run tBTC nodes since the inception. Their past and current contribution is invaluable. They always provided excellent uptime, prompt upgrades, and operational excellence that positively impacted tBTC. Moreover, I cannot count how many times they supported the development team during investigation of various problems affecting the whole systems (e.g. DKG, signing delays, and many more). Last but not least, they did it all without receiving additional gratification from the DAO, despite the DAO provided such gratifications to community beta stakers.
That said, forcibly locking beta stakers and treating them worse than non-beta stakers (who were not providing any actual service to the network) is just unfair. Moreover, such a decision will have a really negative effect on the overall network. Hypothetically, disabling 21% of the nodes will dramatically slow down tBTC operations like deposit sweeps and redemptions for existing wallets (wallets will have less active participants and more retries will be needed to achieve success). Additionally, disabling those nodes will significantly worsen protocol decentralization. Last but not least, locking stakes in such a way will undermine operators’ confidence in the DAO and simply decrease DAO’s credibility in the long term.
I totally get the concerns presented in this proposal. The problem is serious and must be addressed promptly.
I would go even further than what is proposed in this TIP and keep all those nodes operational and participating in key generations for new wallets as long as no transition to BitVM is performed per TIP-100. As long as the network uses 51-of-100 threshold, the operational excellence of existing stakers is crucial as we have already observed heartbeat failures in the past.
I agree the tokens of external stakers should not be locked. They should be replaced by DAO delegations to the same staking provider. It is possible even right now (with caveats) by increasing and later decreasing authorization.